T H E VISCOSITY O F CRUDE-OIL EMULSIONS f

ABSTRACT
A funnel-type viscosimeter suitable for determining When the clean-oil viscosity is known, the viscosity
viscosities of brine-in-oil emulsions was developed and ratio can be used to determine the viscosity of the wet
calibrated against the Saybolt Universal viscosimeter. oil. This information is useful in engineering work, as
Viscosities were determined on "loose," "medium," and the wet-dil viscosity is of importance in designing
"tight" emulsions. The viscosities were largely depen- gathering systems and in choosing pumps and tubing
dent on the amount of brine in the emulsion, and to a sizes for wet wells.
lesser extent on the particle size and age of the
- There is an inversion of phase from brine-in-oil to
emulsion.
oil-in-brine usually in the range from 70 to 85 per cent
The results of the viscosity tests were expressed as
brine content. After the inversion, the viscosity ap-
the "viscosity ratio"; i.e., the ratio of the viscosity of
the emulsion, at any temperature, to the viscosity of proaches that of the brine. From viscosity considera-
the clean oil at the same temperature. tions, the emulsions should be handled with either a
The viscosity ratio at 75 deg
- F for a medium emulsion very low or very high brine content. The practical diffi-
was 1.34, for 10 per cent brine in the emulsion; 1.88 culties of the latter are obvious because of the necessity
for 20 per cent; 4.25 for 40 per cent; 15.0 for 60 per of handling large volumes of fluid with relatively small
cent; and 46.5 for 70 per cent brine. The ratios were volumes of oil.
the same for elevated temperature.

There is little information in the literature regarding There a r e several mathematical formulas f o r calcu-
the viscosities of crude-oil emulsions. Almost all oil- lating the viscosity of a n emulsion, notably the equa-
company laboratories a r e equipped to determine clean- tion developed by Einstein and modified by Smolu-
oil viscosities, and these data a r e usually available. chowski. There is another by Hatschek. None of these
Data a r e presented in this paper which will enable the equations is applicable over a s wide a range of condi-
viscosity of a brine-in-oil type of emulsion to be esti- tions a s found in oil-field emulsions.
mated from the clean-oil viscosity. The brine-in-oil Recently L. T. Monsonl" presented the results of ex-
data a r e of considerable importance in designing field periments to determine the viscosity of brine-in-oil
gathering systems and lines t h a t handle wet oil. Many emulsions. A t t h a t time Monson, in order to express
wells produce a n emulsion, and the viscosity of the his results, introduced the "viscosity ratio," which is
emulsion will influence directly pumping pressures and the ratio of the viscosity of the emulsion a t any tem-
tubing sizes. perature to the viscosity of the clean oil a t the same
An emulsion is a system containing two liquid phases, temperature. The "viscosity ratio" also is used to ex-
one of which is dispersed a s globules in the other. With press results in this paper.
the two liquids, oil and brine, two types of emulsion The previous work by Monson consisted of viscosity
a r e possible: one in which the oil is dispersed in the determinations on brine-in-oil emulsions t h a t were of
brine, and the second in which brine is dispersed in extremely small particle size. The results reported
the oil. Almost all emulsioxis encountered in the oil herein are from determinations on emulsions of both
fields a r e of the latter type; i.e., emulsions in which large and small particle size, which a r e more compara-
brine is dispersed in the oil. Brine is, therefore, the ble to those emulsions met in practice. It was found
dispersed; or internal phase; and the oil is the dis- t h a t the Saybolt viscosimeter was not suitable to de-
persing, or external phase. When considering emulsions termine viscosities on the relatively loose or large par-
in this paper, i t will be of the brine-in-oil type unless ticle-size emulsions because of the small bore of the
otherwise specified. tube, and because some of the brine particles may adhere
The viscosity of a liquid is a measure of the internal to the side of the tube and thereby possibly result in
fluid friction in t h a t liquid; i.e., the resistance to shear, quite large errors in the determination. Another diffi-
agitation, or flow. In addition to the viscosity of the culty with the Saybolt viscosimeter was the settling of
dispersing phase, the viscosity of a n emulsion is com- large particles during tests, which would coalesce to a
plicated by the presence of the dispersed brine particles , layer of brine in the bottom of the Saybolt reservoir.
in the continuous oil phase. The apparatus that was developed to determine these
Prtroleuln Ilectifvlng Co. of Ctllifornia Long Bench Calif.
viscosities consisted of a 2:-in. 60-deg glass funnel with
f Presented n t spring meeting. Paciflc ('bast ~ i s t r i c t , Division
'
of E'roductlon, I,os Angeles, Calif., Jlur. 10. 1912. * I'ig11r1-3 r r f r r t o ltEFEKEN('1CS 011 1). 15:).
a stem length of 6 in. The reservoir t h a t held the emul- during the test. There was no uneven flow because of
sion contained a paddle f o r stirring, and thereby main- separation of the phases of the emulsion, nor did brine
t a i n d a uniform m ~ x t u r eflowing through the funnel. adhere to the inside of the funnel during ,any of the
A n overflow tube was placed in a catch-all below th8 tests.
funnel to drain off excess emulsions. All the tests made F o r comparison with other work it was necessary t o
with the funnel viscosimeter were made a t room tern- cori.elate efflux times obtained on the funnel viscosim-
perature. However, the equipment could be surrounded ' eter with Saybolt Universal seconds. Several viscosity
by a water jacket and used in making determinations deternilnations were carried out by both methods on
a t elevated temperatures. transformer oil, glycerin, and on samples of E-24
The upper vessel was filled with t h e emulsion to be (Goose Creek, Texas) dry crude, E-448 (Mt. Poso,
Calif.) crude, and E-619 (Seal Beach, Calif.) dehy-

R- THERMOME TER
' drated crude. The test data, which a r e outlined in
Table 1, when corrected to the same teinperature f o r
each series of determinations, give a correlation curve
f o r converting funnel, viscos~tyto Saybolt Universal vis-
cosity, which is presented in Fig. 1.
After correcting the observed viscosities to the same
temperature by means of the ASTM viscosity-tempera-
-1000 ML. WOULFF BOTTLE

1 1 /*oraron 30 RPM TABLE 1

Calibration of Funnel Viscosimeter
Vlscnsitv I))
RUBBER TUBING Sarbolt Ylscosltr 1,r
LTnlrarsnl Ternpera. Funnel Tempera-
PINCH COCK Viscosiiiieter ture V ~scosimeter ture
Material (Seconds) , (De:: F) (Secoi~<ls) (Dez F)
E-448 crude, 16.2
deg A P I . . . . 10,135
Glycerin . . . . . . . ' 1,740
E-24 crude, 25.1
deg A P I . . . . . 701
E-619 crude, 25.1
deg A P I . . . . 255
E-619 crude, 25.3
deg A P I . . . . 202
Transformer oil. 90 '
Water . . . . . . . . . 30.5

ture chart, the curve f o r Saybolt Universal viscosity vs.

funnel viscosity is a straight line, a s shown in Fig. 2,
with the formula :
Saybolt viscosity = (funnel viscosity minus 2.5) (5.41)
Apparatus for Determining Viscosities of Crude-Oil
En~ulsions. Effect of Time
FIG. 1 ' The time that elapses between the formation of the
emulsion and the determination of its viscosity will
tested and a g ~ t a t i o nwas started. After the oil had been affect the results obtained. I n general, aging will in-
allowed to reach equilibrium a t room temperature the crease the viscosity, especially of medium and loose
test procedure consisted in allowing the oil to flow emulsions. The effect is relatively small, and actual
through the filling tube and through the funnel. The measure~nentsto determine this effect were not made.
pinch-clamp was adjusted so t h a t the funnel, which was However, all saniples were aged 15 mi11 before testing
held level, was kept full and slightly overflowing a t all in order to make results comparable.
times. The 'efflux t ~ m ewas nieasured a s the time re-
quired f o r 100 in1 of sample to flow through the funnel. Effect of Particle Size
The teinperature was measured by a thermometer sus-
pended in the funnel chamber. This method of vis- The effect of particle size on the viscosity of a n emul-
cosity determination gave reproducible results, efflux sion 1s larger in the higher range of brine content, and
times sufficiently large to be accurate, but not so long tests were made to determine the viscosity of eniulsions
a s to introduce the possibility of temperature variations having loose, medium, and tight characteristics. This
nomenclature is only qualitative, and centrifuge tests emulsions of different characteristics were used in these
were used to determine the approsinlate particle s u e tests. "Tight" emulsions a r e those of extremely small
rather than photomicrographs, a s centrifuge tests .are pal-ticle size, and 'usually occur when a wet well is
more suitable f o r field use. A centrifuge cut was made flowing with a high gas-oil ratio and IS subjected to
on each emulsion by mixing 25 ml of the emulsion with extreme conditions of agitation. A cut made on a n
emulsion produced under these conditiohs would show
all emulsion. A "loose" emulsion is one which nor-
mzlly occurs in the production from a punlping well in
which mechanical equipment is in good condit~on.About
30 per cent of the material centrifuged out when inak-
ing a cut will consist of emulsion and about 70 per cent
wlll be brine. A "medium" emulsion is between a "tight"

FUNNEL VISCOSITY, SECONDS

Calibration of Funnel Viscosimeter to Saybolt Universal
Viscosimeter.
FIG. 2

I Type
TABLE 2
Characteristics of En~ulsions
Appearance of Cut
II Brine with Brine Content.
FIG. 3
~ s Crude Oil and
Variation in Viscosity of E n ~ u l s i o ~of

Tight emulsion All emulsion

Medium emulsion 80 per cent emulsion, and "loose" en~ulsion,and may occur under various con-
20 per cent brine ditions such a s a pump in poor condition o r leaky tub- ,
Loose emulsion 30 per cent emulsion, ing, o r g a s coming through the tubing with t h e wet
70 per cent brine oil. I n this case about 80 per cent of the cut will be
emulsion and 20 per cent wlll be brine.
75 ml of gasoline-carbondisulfide * and centrifuging for
10 min a t room temperature a t a speed of 1,400 rpm. I The effect of the particle size is shown in Flg. 3, to-
gether with the effect of the amount of brine.
The emulsification was controlled in order to give emul-
sions of approximately t h e characteristics by centrifuge
shown in Table 2.
I The Effect of Brine Content
As i t was impossible to obtain natural-occurring
Emulsions with different degrees of emulsification
emulsions of varying particle size which also contained
occur in the .production of crude oil; and, therefore,
a varying per of brine, it was necessary to make
* 25 per cent g n s o l i ~ i e ,7.5 llrr ceut c a r b o ~ ~ d ~ s u l h d e . these en~ulsionsby mixing the oil and brine intimately
 i n a piston-type emulsifier. This emulsifier consisted of t e n t of 70 per cent; but, owing to t h e e-xtremely high
- a cylinder with a loose-fitting piston. The brine a n d oil viscosity of the emulsion, i t would not flow through the
were poured into t h e cylinder, and agitated by moving viscosimeter, and the test was abandoned. I t w a s pos-
'
the piston up and down rapidly by hand. Emulsions sible t o deterinine t h e viscosity of t h e medium emulsion.
were made in this manner contaimng u p to 70 per cent a t 70 per cent, which was 46.5 times t h a t of t h e clean
of brlne. These einulsions were made from a Seal oil. Difficulty also was encountered in t r y i n g t o deter-
Beach crude, with a gravity of 25.1 deg A P I a t 60 deg mine t h e viscosity of t h e loose emulsion with 70 per
F, and brine produced with this crude which had a cent brine content because of stratification of the test

TABLE .3
Test Data on Loose En~ulsions
P e r Cent Brine Added
5 - 10 20 40 60
'c
Emulsion, per c e n t . . ............................... 1.4 2.9 10.0 14.0 22.0
Brine, per c e n t . . ......... :. ........................ 3.7 7.7 12.0 26.8 45.6

Total, per c e n t . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1 10.6 22.0 40.8 67.6

Temperature of test, deg F. ........................ 73.5 73.5 72.5 71 70
Observed funnel viscosity, s e c . . ................... ., 53 61 85 201 597
Calculated Saybolt Universal viscosity, sec.. .......... 270 314 440 1,070 3,220
Calculated Saybolt Universal viscosity, a t 75 deg F, sec. 261 - 301 410 915 2,610
Viscosity ratio a t 75 deg F , calculated.. ............. 1.16 1.34 1.82 4.06 11.6

TABLE 4
Test Data 011 Medium El~lulsions
P e r Cent Brine Added

En~ulsion,per cent.. ................................

B r i n e , p e r cent .....................................

Total, per cent. . . ; ...................... ;......

Temperature of test, deg F . . ........................
Observed funnel viscostty, s e c . . .....................
Calculated Saybolt Universal viscosity, sec. . . . . . . . . . . .
Calculated Saybolt Universal viscosity a t 75 deg F, sec.
Viscosity ratio a t 75 deg F , calculated.. ..............

specific gravity of 1.023 a t 68 deg F. The viscosity sample. I11 other words, under the conditions of the
characteristics of this crude were: experiment, it was not possible to hold 70 per cent brine
uniformly dispersed in this emulsion.
Deg F Seconds There is little difference i n the viscosity ratio f o r
Viscosity (Saybolt Universal) a t . . 70.5 255 brine contents up t o 30 per cent. Above this value t h e
Viscosity (Saybolt Universal) a t . . 100 127 curves d ~ v e r g e ,and the effect of particle size becomes
Viscos~ty(Saybolt Universal) a t . . 119 9d of importance.
,Viscosity (funnel) a t . . .......... 71 50 Fig. 4 shows a generalized curve of the variation in
relative viscosity with increase of brine content. With
The viscosity deterininations on t h e samples of emul- no brine present, t h e viscosity 1s t h a t of the clean oil.
sions prepared a s described above a r e tabulated i n As t h e brine content increases, the viscosity increases
Tables 3, 4 , and 5. gradually until t h e emulsion is about 40 per cent brine,
These d a t a a r e shown in Fig. 3, which .gives t h e and then r a t h e r abruptly increases until t h e brine con-
change in t h e viscosity of the emulsion f o r variations In t e n t reaches 70 o r 80 per cent. Generally, there will be
brine content. Curve A of Fig. 3 i s t h e curve f o r loose a n inyersion of, phase in this region, and t h e brine will
emulsion; curve B f o r medium emulsion; and curve C become the external phase. The internal phase will be
f o r tight emulsion. a brine-in-oil einulsion. I n this case a so-called "double" '

The highest brine content in the tight-emulsion curve enlulsion i s fonned, i n which a brine-in-oil emulsion i s
is 60 per cent. A test was attempted with a brine con- dispersed a s t h e internal phase to form a n oil-in-brine
 emulsion. When an en~ulsionof this type flows into a was considered a loose emulsion. The data from these
production tank, the brine in, the external phase settles tests a r e given in Table 6, and a r e plotted In Fig. 5,
rapidly, leaving a normal brine-in-oil emulsion. together with the viscosity ratio of loose emulsion made
After inversion to the oil-in-brine type of emulsion, from the 25-deg-API oil made a t 75 deg F.
the viscoslty wlll drop to a polnt slightly higher than Results of tests made a t various temperatures on
the viscosity of the brine and gradually decrease u n b l naturally-occurring emulsions a r e given in Table 7.
we have the viscosity of 100 per cent brine. The point From the ratios given in Fig. 5, and also in Table 6,
a t which inversion from a brine-ln-oil to a n 011-in-brine i t is apparent that the viscosity ratio for a given brine
emulsion takes place depends on the character of oil, content is independent of the temperature.
character of brlne, and kind and degree of emulsific?- Monson determined the viscosity of emulsions a t va-
tion. I n general, this point may be expected a t 70 to 85 rious temperatures, and plotted the results on ASTM
per cent brine content. However, In spec~ficcases there charts f o r viscosity-temperature relationship. The vis-
may be a n inversion a t a s low a brine content a s 50 per coslty-temperature curve for a n emulsion, a s glven by
cent, and in other cases brine-in-oil emulsions a r e Monson, was a straight line, and approsimately parallel
i
stable wlth a s much a s 95 per cent of brine dispersed to the curve f o r the clean 011. .
in 5 per cent of oil. When emulsions near the inversion
polnt a r e approached, the viscosity becomes very high,
and emulsions a t this point can be considered a s plastic
1 Conclusions
solids, especially ~f brine 1s very finely,divided o r a An apparatus was developed suitable f o r determining
very tight emulsion. Very few determinations have viscosit~esof crude-oil emulsions, and calibrated against
been made on emulsions containing more than 60 to 70 a Saybolt Universal viscosimeter.

TABLE 5
Test Data on Tight E~nulsions
P e r Cent Brine Added
5
Emulsion, per cent. .................................. 6.4
I B r i n e , p e r c e n t .......................................

. Total, per cent.. ............................... 6.4

Temperature of test, deg F . . ........................ 80 ...
Observed funnel viscosity, sec.. ...................... 40 Infinity
Calculated Saybolt Universal viscosity, see. ........... 203
Calculated Saybolt Universal viscosity a t 75 deg F , see. 235
. Viscosity ratio a t 75 deg F, calculated.. .............. 1.20
0

per cent brine because of the difficulties experienced The viscosity of a n emulsion is largely dependent on
when working with emulsions with these high brine the ainount of brine in the,emulsion and, to, a lesser
contents. extent, on the particle size and age of the emulsion.
From a n inspection of Fig. 4, i t is apparent t h a t
emulsions should be pumped either wlth a low brlne
co1:tent a s a brine-in-oil emulsion, or with a n extremely
high brine content a s a n oil-in-brine emulsion. The dis-
advantage, from a practical standpoint, in pumping
. oil-in-brlne enlulsions is the large volume of fluid t h a t
it is necessary to handle. This, however, is being done
in a t least one case, where a low-gravity wet crude
is pumped cold through a line, by injecting brine Into t h e
line with the wet oil to reduce the pressure a t the pump.

Effect of Temperature
A few tests were made several years ago using a
Saybolt Universal viscosimeter to determine the effect of
temperature on the viscoslty of a low-gravity ccude-oil
emulsion, made from a crude 011 with a gravity of 14.8
deg API. A low-gravity crude was used to eliminate General Curve for Viscosity of an Emulsior~with 0 to 100
changes in viscosity due to evaporation of some of the Per Cent Brine (Medium Emulsion). ,
light fractions;from the crude. The emulsion formed FIG. 4
 The viscosity ratio, which is the ratio of the viscosity
of the emulsion to the viscosity of the clean oil a t a
given temperature, can be used to estimate the vis-
cosity of the emulsion for engineering work.
A general curve was given to show the change in
viscosity of a n emulsion from 0 per cent to 100 per
cent brine. From a consideration of this curve, it is
apparent t h a t crude-oil emulsions should be pumped
with either a low brine content o r a n extremely high
brine content in order to maintain low pumping pres-
sures.

ACKNOWLEDGMENT
I n closing, the author wishes to acknowledge the help
given him by members of the Petroleum Rectifying
Company laboratory staff-especially to Lieut. D. B.
Smith, now with the Chemical Warfare Service, and
to C. S. Stevens, who made most of the tests under
the direction of C. H. M. Roberts, research director;
also to the Petroleum Rectifying Company of Cali-
fornia for permission to publish these results.
Effect of Temperature o n Viscosity Ratio.
REFERENCES
FIG. 5 I

1 Louis T. &Ionson, I n d . Eng. Chem. 30, 1287 (1938).

TABLE 6
Viscosities at Elevated Temperatures DISCUSSION
Per Cent Brine J. E. Sherborne (Union Oil Company of California,
Santa F e Springs, Calif.) : The data presented by Mr.
Woelflin a r e of considerable interest to those who al-
Temperature, deg F. .. . 180 most daily attempt to determine pressure drop result-
Viscosity (Saybolt Uni- ing from flow of wet crudes through various units of
. . . 170
versal), sec . . . . . . field equipment. I n the past it has been necessary to
.
Viscosity ratio . . . . . . . ... make individual tests on nearly every system before
Temperature, deg F . . . . 140 accurate information a s to the proper size a n d kind
Viscosity (Saybolt Uni- of equipment for a given installation could be made.
versal), sec . . . . . . . . . 470 With such data a s presented in Fig. 3, it is possible
.
Viscosity ratio . . . . . . . ... to design the proper size of pipe lines, heat interchang-
ers, and related equipment without the necessity of
TABLE 7 testing the particular type of crude to be handled; be-
Viscosities of Natural Emulsiolls cause, a s Mr. Woelflin says, the viscosities of the d r y
crudes a r e nearly always available.
Round Round It is of interest to note t h a t tests made in the Union
Kern Moun- Moun- Oil Company laboratory, with a different type of vis-
Source Front tain tain cosimeter-one consisting of a source of wet crude un-
Gravity of crude, deg A P I . .
15.8 14.6 16.3 der pressure, a coil of copper tubing immersed in a con-
Brine, per cent.. . . . . . . . .. .
9.0 18.0 20.0 stant temperature bath, and pressure gages to measure
Viscosity ratio a t 120 deg F. 1.22 1.55 1.50 the pressure differential within the copper tubing-
Viscosity ratio a t 140 deg F .
1.22 1.60 1.63 gave results which fit the curves of Fig. 3 within the
Viscosity ratio a t 160 deg F . 1.23 1.58 1.63 limits of the data presented by Mr. Woelflin.